The invention relates to a method for the production of picture sequences for a fast search operation in a film playback device. So-called Video-on-demand services (VOD) are currently being developed, in the case of which the user of a television set is coupled via a return channel to a provider of Video-on-demand services. By this means he can request selected broadcasts and/or films. The selected film is then transmitted from the VOD service provider to the subscriber via a corresponding data transmission channel.
The user of this service also frequently requires so-called trick modes. These trick modes relate to a fast search in the forward direction and in the backward direction, a pause function as well as a slow-motion function. These trick modes are also known from conventional video recorders. The request for a trick mode is then effected, for example, using the remote control of the television set and is transmitted via the return channel to the provider of the VOD service. While the trick modes pause and slow motion are relatively easy to implement, for this it being necessary to transmit the last decoded picture continually or repeated a number of times, the trick modes fast forward run and reverse run are more difficult to implement.
The object of the invention is to make such trick mode services available in the case of VOD service providers and film playback devices. This object is achieved by claim 1. The VOD service providers use so-called video servers for making the selectable programmes and films available. A video server essentially consists of one or a number of hard disks on which the video data are stored. For this purpose, the video data are, for example, digitally coded according to the MPEG Standard (Moving Picture Experts Group) MPEG1 or MPEG2 known from the prior art.
In order to achieve the necessary high data compression, the individual pictures of a sequence are coded differently in accordance with the MPEG2 Standard. A distinction is made between intra-coded pictures (I frames), unidirectionally predicted pictures (P frames) and bidirectionally predicted pictures (B frames).
However, the intra-coded pictures require a greater storage capacity on the hard disk than the unidirectionally or bidirectionally predicted pictures. The intra-coded pictures can be decoded independently of preceding or succeeding pictures, but the preceding intra-coded or unidirectionally predicted picture is required for the decoding of unidirectionally predicted pictures. Both the preceding intra-coded or the unidirectionally predicted picture and the succeeding intra-coded or unidirectionally predicted picture are required for the bidirectionally predicted pictures. The sequence of the I-, P- and B-coded pictures is freely adjustable to a large extent in accordance with the MPEG1 or MPEG2 Standard. However, it has been shown in the past that a scheme is advantageous in which three unidirectionally coded pictures follow an intra-coded picture and two bidirectionally coded pictures should be stored between in each case two unidirectionally coded pictures and between an intra-coded picture and a unidirectionally coded picture. This sequence is illustrated in FIG. 7. Such a sequence comprising 12 frames is constantly repeated during recording, typically approximately every 0.5 seconds.
In order to realize a fast forward run in the case of pictures coded in accordance with the MPEG Standard, a technique such as is known from analog video recorders can no longer be applied, on account of the great data compression. A possible solution in which, for example, an increase in the playback speed by a factor of 12 would be achieved by transmitting only the intra-coded pictures to the user also encounters difficulties, since most of the bits when coding video sequences are required for coding the intra-coded pictures. If, therefore, it was desired to transmit only the intra-coded pictures in the case of such a fast forward run, then it is no longer possible with the available broadband cables and desired numbers of channels, on account of the number of bits per picture, where the number of bits is far above the average in the case of these pictures.
Another possible method would be to use a second bit stream as well as the bit stream, produced in accordance with the MPEG Standard, for the regular playback speed, which second bit stream has been obtained from a sequence which has already been subsampled in the time domain, and additionally to store the said second bit stream on the storage medium. A switch could then be made to this bit stream in the event of a desire for a fast forward run. However, this method would have the disadvantage that additional storage space would be used up for a service which would probably be used relatively rarely.